Why is Windows sample rate still not dynamic ?

[Berwhale]

"over", so your music would play in native sampling rate and the sounds would get resampled. Is that bad?

Is that any better than setting the Windows sample rate to 96KHz and bit depth to 24 and forgetting about it?

 

[Jose Hidalgo]

If you want bit-perfect operation, use WASAPI Exclusive, ASIO, or WDM/KS.

I certainly DON'T want bit-perfect operation ! Never said I wanted that.

• First, bit-perfect is a myth and we should move past it now (see the link below).
  
• I'm equalizing all my audio devices. It's WAY better than dreaming of "bit-perfect".
• For that purpose I'm using Equalizer APO, plus the audio app that I've developed and which you can find in my signature.

What sample rate would you suggest that Windows should choose if two pieces of audio are playing at the same time and those two pieces of audio have different sample rates?

That is of course an extreme case. But even that could be managed by resampling. Windows could very well choose the sample rate of the first playing audio, then when the second comes in, it would be automatically resampled to the samplerate of the first one. Problem solved. Would that create unnecessary resampling ? Yes, sometimes, and only for people playing several audio streams at the same time (which isn't exactly a majority of people). But for all people wanting to play ONE SINGLE STREAM OF MUSIC (and that's definitely a majority of people), that would be a huge plus.

The short answer is, set the rate to the highest values your DAC or audio interface supports. That will provide the best overall experience

I'm not convinced of that, unless measurements say otherwise. Why resample 98% of my files ? I would rather set everything to 32/44.1, like I'm already doing. No resampling + overhead for digital volume control.

If you really must have bit-perfect output

What do you guys have with bit-perfect these days ? Let's all move past it !

3. For playing media I recommend WASAPI Exclusive mode.

DDF already proved beyond any reasonable doubt that it wasn't necessary. It has been a long standing myth. I'm glad it's over now.

 

[edechamps]

I certainly DON'T want bit-perfect operation !

If you don't care about bit-perfect operation (and I definitely agree you shouldn't care - I don't either), then why are you obsessing over benign sample rate conversions that you won't hear anyway?

 

[Jose Hidalgo]

If you don't care about bit-perfect operation (and I definitely agree you shouldn't care - I don't either), then why are you obsessing over benign sample rate conversions that you won't hear anyway?

Hi Eric (it's Eric, right ?). I don't care about bit-perfect because I'm EQing sound. If I wasn't, I think I would care. Actually I have cared about bit-perfect for many years before giving in and entering the wonderful EQ world.

I never heard that resampling was 'benign'. Is it really ? If it was, we wouldn't even discuss about it. But on the contrary, there's a lot of litterature on that subject and many comparisons between different resampling algorithms. Plus I'm willing to bet that Windows resampling algorithm isn't exactly among the best ones.

As for being more or less audible, I think that's not the subject. In that case Amir wouldn't be comparing DACs that are all above 110 dB SINAD, which is beyond human hearing ability, and he would have a lot less work.

 

[danadam]

Is that any better than setting the Windows sample rate to 96KHz and bit depth to 24 and forgetting about it?

In general, for me, no.
But in the context of this thread, for people who for some reason want to avoid resampling, I think yes.

 

[edechamps]

I never heard that resampling was 'benign'. Is it really ? If it was, we wouldn't even discuss about it. But on the contrary, there's a lot of litterature on that subject and many comparisons between different resampling algorithms. Plus I'm willing to bet that Windows resampling algorithm isn't exactly among the best ones.

Yes, it is benign. If that's news to you then you've been spending too much time on audiophile forums

It's easy to demonstrate, too. Just switch the audio device settings between the two sample rates and compare with and without Windows resampling. You can even do a more rigorous test yourself: resample some music file using some high-quality offline resampler, then compare the original and resampled files (one of which will necessarily get resampled by Windows on playback) using the foobar2000 ABX comparator or equivalent.

I'm willing to bet you will fail that ABX test. I certainly would.

In that case Amir wouldn't be comparing DACs that are all above 110 dB SINAD, which is beyond human hearing ability, and he would have a lot less work.

Well you know, this is one of the traps of Amir's electronics reviews: he's so demanding that once you've read a few of them you start losing perspective and start demanding 110 dB SINAD while forgetting that it's extremely hard to hear any difference above 80 (and that's generous - some would even say 60). Amir is looking for engineering perfection, not practical transparency, which most electronics have already reached a long time ago.

 

[Promit]

I never heard that resampling was 'benign'. Is it really ? If it was, we wouldn't even discuss about it. But on the contrary, there's a lot of litterature on that subject and many comparisons between different resampling algorithms. Plus I'm willing to bet that Windows resampling algorithm isn't exactly among the best ones.

Tsk, tsk. Where do you think we are? This community has already been over this -
Windows resampling not actually that bad? | Audio Science Review (ASR) Forum
tl;dr the errors introduced by recent Windows versions are way down in amplitude, as long as you keep overall volume below 95% or so due to some anti-clipping functionality (that can be bypassed - read links at end of thread). Win7 might not be such a safe bet, though.

 

[Marc v E]

What sample rate would you suggest that Windows should choose if two pieces of audio are playing at the same time and those two pieces of audio have different sample rates?

I could be wrong, but isn't it all to do with focus and easily accessable preferences?
For example: if the audio player is on in my OS I expect it to have top focus. It should play correctly, or at least I should be able to configure it to behave in that way.

It reminds me of the shifting focus on windows when it's starting up the os and associated programs. If the home screen shows and I open my mail the next program that is automatically started gets focus instead of the mail I just clicked. Weird and counterintuitive imo.
Mac Os is better is this regard.

I must say I really don't value the ability of an OS to play multiple sounds from multiple sources at the same time. OS sounds are more like machine f@rts than anything useful to me.

 

[Promit]

I could be wrong, but isn't it all to do with focus and easily accessable preferences?
For example: if the audio player is on in my OS I expect it to have top focus. It should play correctly, or at least I should be able to configure it to behave in that way.

For me, which app has focus (a UI property) has absolutely !@#$-all to do with which apps are creating audio that I care about and therefore require audio priority. (And you would very much not want your audio subsystem to track the UI state.) Now if you mean that the audio app should have processing priority, well, yes. That's been solved for many, many years. But the audio subsystem has no concept of "important" vs "unimportant" audio. It does have a concept of exclusive mode, which seems to apply to some of the users in this thread. With the correct audio player software, you can indeed hand over control of your audio stack entirely to that one application and not worry about system sounds and other sources.

 

[Marc v E]

For me, which app has focus (a UI property) has absolutely !@#$-all to do with which apps are creating audio that I care about and therefore require audio priority. (And you would very much not want your audio subsystem to track the UI state.) Now if you mean that the audio app should have processing priority, well, yes. That's been solved for many, many years. But the audio subsystem has no concept of "important" vs "unimportant" audio. It does have a concept of exclusive mode, which seems to apply to some of the users in this thread. With the correct audio player software, you can indeed hand over control of your audio stack entirely to that one application and not worry about system sounds and other sources.

Well yes, priority is what I mean.

 

[Offler]

DDF already proved beyond any reasonable doubt that it wasn't necessary. It has been a long standing myth. I'm glad it's over now.

I did my own research on that topic (not watching youtube, but actually measuring things )
https://www.audiosciencereview.com/...indows-audio-quality-debate.19438/post-807219

I still recommend Wasapi Exclusive for CDs, FLACs, DVDs and BluRays.

Also i use the tweaks to avoid CAudioLimiter, mainly for gaming audio where no other options are available.

 

[Pennyless Audiophile]

You may find this https://www.thewelltemperedcomputer.com/SW/Windows/Win10.htm article interesting

macOS core audio has similar issue with fixed sample rates, at least in some configurations I've encountered.

I have a 404 error when I click the link...

 

[krumpol]

It's been copy-pasted poorly, here you go:
https://www.thewelltemperedcomputer.com/SW/Windows/Win10.htm

 

[Katji]

Tsk, tsk. Where do you think we are? This community has already been over this -
Windows resampling not actually that bad? | Audio Science Review (ASR) Forum
tl;dr the errors introduced by recent Windows versions are way down in amplitude, as long as you keep overall volume below 95% or so due to some anti-clipping functionality (that can be bypassed - read links at end of thread). Win7 might not be such a safe bet, though.

8.1 ?
Whatever I've read [at microsoft.com], I never quite got it clear, it seemed that the major/significant changes were done in 8.1, before 10.

 

[Offler]

8.1 ?
Whatever I've read [at microsoft.com], I never quite got it clear, it seemed that the major/significant changes were done in 8.1, before 10.

Windows audio stack significantly changed in Windows Vista. That model is still being used today.

 

[Katji]

ok...

That does not make it worthwhile to me to change to Windows 10. (And given what nonsense Windows 11 appears to be, if I eventually can't continue with 8.1 [install on new laptop], then Linux is in my future.)

flac ... Windows 8.1 and WMP plays flac, and whatever flac I play via web pages /Firefox. [Just not supported by 8.1 Windows Explorer.]

Irrelevant to me. I would still use the DAC manufacturers' Thesycon/XMOS ASIO driver. [I just need to know before I get a DAC. So far so good.]

[...]

 

[dualazmak]

Just for your reference, I have "historical" and "practical" reasons for sticking to "all in ASIO I/O" in my PC based multichannel audio system, as I posted here.

 

[Katji]

Just for your reference, I have "historical" and "practical" reasons for sticking to "all in ASIO I/O" in my PC based multichannel audio system, as I posted here.

Noted. It relates to my point/s...

- that there is no benefit or no worthwhile improvement in audio from Windows 8.1 to 10, and audiophile articles/pages often/generally refer to Windows 7 or XP to Windows 10 (as though Windows 8.1 never existed)

- and flac plays on 8.1

- Windows 10 is more "bloated" (as per usual, and adds no functionality - at least not anything that affects me) and has a history of updates causing trouble.

But I have no use for ASIO4ALL generic, I need whatever the DAC manufacturer implements in the XMOS/Thesycon driver.

 

[Jose Hidalgo]

Yes, it is benign. If that's news to you then you've been spending too much time on audiophile forums

Haha, I think I've never been to such dark and cursed places.

Tsk, tsk. Where do you think we are? This community has already been over this [...] Win7 might not be such a safe bet, though.

Well, I still have Win 7 in my main rig, and I won't upgrade it because I'm going to replace my rig in the next few months (when latest-gen laptops FINALLY become available ). So I guess I was right after all in assuming that Windows wasn't such a safe bet.

I did my own research on that topic (not watching youtube, but actually measuring things )
https://www.audiosciencereview.com/...indows-audio-quality-debate.19438/post-807219

I still recommend Wasapi Exclusive for CDs, FLACs, DVDs and BluRays.

Also i use the tweaks to avoid CAudioLimiter, mainly for gaming audio where no other options are available.

I've seen your post, but what does it prove exactly ? (besides that 0.12 dB treshold). And why would you recomment Wasapi when DDF proved that it wasn't necessary ? For the record, my foobar output is never at 0 dB. It's somewhere in the -5 dB to -20 dB range I guess.

Plus like wisely said in the topic you mentioned... https://www.audiosciencereview.com/...mpling-not-actually-that-bad.9092/post-385960

 

[Offler]

I've seen your post, but what does it prove exactly ? (besides that 0.12 dB treshold). And why would you recomment Wasapi when DDF proved that it wasn't necessary ? For the record, my foobar output is never at 0 dB. It's somewhere in the -5 dB to -20 dB range I guess.

Plus like wisely said here... https://www.audiosciencereview.com/...mpling-not-actually-that-bad.9092/post-385960

It proves that on -0,14dB there is no added distortion and software measured -80dB as is expected from my amplifer.
At -0,12dB there is some added distortion, roughly -60dB average when CAudioLimiter was triggered.

Just keep in mind this value is measured on my specific setup, and might be different for yours.

I recommend WASAPI Exclusive mode because it prioritizes audio stream and completely bypasses rest of windows audio stack. Ofc it bypasses CAudioLimiter, but that is not my point.

For example when I watch movie, no other sound will beep from the background, because MediaPlayer took the device all for itself.

There will be less jitter (if any). and this experience comes from time when I used standard desktop PC to feed old 8bit computer with programs in MP3 or WAV format.

(Actually I wanted to test this again just for fun . It was easy for a dedicated CD player to match 3600baud rate without any error, but it was much an issue for a software.)